The F-statistic is 93.38 for the linear model and 96.28 for the log-linear model, with both having a p-value less than 0.0001 . This indicates that the overall regression model is statistically significant and that the explanatory variables collectively have a significant effect on the dependent variable (wages), strengthening the case for including IQ in the wage model .

The p-value for the coefficient of IQ is 0.000, which is highly significant . This means that there is strong evidence against the null hypothesis that the coefficient is zero (i.e., there is no relationship between IQ and wages). Therefore, we can be confident that the relationship observed between IQ and wages is not due to random chance .

The regression analysis indicates that IQ has a positive effect on wages, with a coefficient of 1.65108. This means that each additional point in IQ is associated with an increase of approximately $1.65 in wages . The R-squared value of 0.0941 suggests that approximately 9.41% of the variation in wages can be explained by IQ alone. This relatively low R-squared indicates that while there is a statistically significant relationship between IQ and wages, most of the wage variation is due to other factors not captured in the model .

The minimum IQ score is 50, and the maximum is 145, reflecting a wide range of cognitive abilities within the dataset . The range of wages, from approximately $1023 to $1615.6, represents considerable economic diversity, suggesting varying levels of job roles and industries among the individuals . The substantial variation in both IQ scores and wages implies a heterogeneous population where different levels of cognitive abilities potentially lead to different economic outcomes, which is valuable for understanding socioeconomic dynamics.

The adjusted R-squared value is 0.0931 for the linear model and 0.0957 for the log-linear model . These values indicate that the explanatory power of the model is slightly improved in the log-linear version, suggesting that log-transformation may fit the data marginally better. However, both models explain only a small proportion of the variability in wages, indicating that other factors not included in the model are likely significant in predicting wage outcomes .

In the log-linear model, the coefficient of IQ is 0.0013599 . This implies that for each additional point in IQ, the wage increases by approximately 0.13599% . In a log-linear context, coefficients can be interpreted as approximate percentages, assuming the change is small, which fits our case here.

The log transformation is important because wage distributions are typically right-skewed, and log transformation helps to normalize the data, stabilizing variance and allowing the model to meet linear regression assumptions more closely . This transformation changes the interpretation from an additive effect to a multiplicative effect, such that the coefficient now represents a percentage change rather than an absolute change in wage for each unit increase in IQ .

The root mean squared error (RMSE) is 77.235 for the linear model and 0.06265 for the log-linear model . A lower RMSE in the log-linear model suggests a better fit, indicating that predictions of the natural log of wages are closer to the actual data points than the predictions in the linear model. The magnitude of RMSE also informs us of the average distance between observed and predicted values, implying better precision in the log-transformed model .

The standard error for the IQ coefficient in the linear model is 0.1708609, and for the log-linear model, it is 0.0001386 . The relatively small standard errors suggest that the coefficient estimates are precise, indicating a consistently measurable effect of IQ on wages across the sampled population. The small standard error in the context of the log-linear model particularly indicates that even changes as minor as 0.14% due to IQ alterations are estimated with certainty .

In the linear regression, the constant term is 1024.364, implying that when IQ is zero (a hypothetical scenario), the wage would be expected to be approximately $1024.36 . In the log-linear model, the constant term is 6.943094, which represents the natural log of the expected wage when IQ is zero, corresponding to an expected wage of about 1034.21 dollars after exponentiation . Both interpretations are theoretical as IQ scores can't realistically be zero.